The camera internal parameters calibration method based on building the initial measurement of the network is put forward. The lack of direction depth of the plane target, as well as the constraints in three-dimensional target space, is solved effectively. In accordance with the principle of measurement network, after the imaging of the target board, the initial measurement network is established. After resection for solving the exterior orientation, forward intersection for the spatial coordinates of encoded points attached on the target board, the camera parameters are solved by the iterative optimization finally. And then, the internal parameters are used for solving the space coordinates. The experimental results show that an average error of building the initial network value of 0.0794 is superior to values of -0.2443 and -0.1916 of planar-taget and three-dimensional-taget calibration. Besides, the calibration time is significantly less than the virtual three-dimensional calibration. This method has the advantages of fast, accurate and efficient, can meet the requirements of the current large-scale vision measurement in the internal parameters calibration on site.